Merge branch 'r25.10' of github.com:triton-inference-server/server into yinggeh/tri-49-request-for-openai-compatible-api-endpoints-for-triton

Author: yinggeh

Dockerfile.sdk

@@ -29,7 +29,7 @@
 #
 
 # Base image on the minimum Triton container
-ARG BASE_IMAGE=nvcr.io/nvidia/tritonserver:25.09-py3-min
+ARG BASE_IMAGE=nvcr.io/nvidia/tritonserver:25.10-py3-min
 
 ARG TRITON_CLIENT_REPO_SUBDIR=clientrepo
 ARG TRITON_REPO_ORGANIZATION=http://github.com/triton-inference-server

README.md

@@ -29,8 +29,8 @@
 
 >[!WARNING]
 >You are currently on the `main` branch which tracks under-development progress
->towards the next release. The current release is version [2.61.0](https://github.com/triton-inference-server/server/releases/latest)
->and corresponds to the 25.09 container release on NVIDIA GPU Cloud (NGC).
+>towards the next release. The current release is version [2.62.0](https://github.com/triton-inference-server/server/releases/latest)
+>and corresponds to the 25.10 container release on NVIDIA GPU Cloud (NGC).
 
 # Triton Inference Server
 
@@ -90,16 +90,16 @@ Inference Server with the
 
 ```bash
 # Step 1: Create the example model repository
-git clone -b r25.09 https://github.com/triton-inference-server/server.git
+git clone -b r25.10 https://github.com/triton-inference-server/server.git
 cd server/docs/examples
 ./fetch_models.sh
 
 # Step 2: Launch triton from the NGC Triton container
-docker run --gpus=1 --rm --net=host -v ${PWD}/model_repository:/models nvcr.io/nvidia/tritonserver:25.09-py3 tritonserver --model-repository=/models --model-control-mode explicit --load-model densenet_onnx
+docker run --gpus=1 --rm --net=host -v ${PWD}/model_repository:/models nvcr.io/nvidia/tritonserver:25.10-py3 tritonserver --model-repository=/models --model-control-mode explicit --load-model densenet_onnx
 
 # Step 3: Sending an Inference Request
 # In a separate console, launch the image_client example from the NGC Triton SDK container
-docker run -it --rm --net=host nvcr.io/nvidia/tritonserver:25.09-py3-sdk /workspace/install/bin/image_client -m densenet_onnx -c 3 -s INCEPTION /workspace/images/mug.jpg
+docker run -it --rm --net=host nvcr.io/nvidia/tritonserver:25.10-py3-sdk /workspace/install/bin/image_client -m densenet_onnx -c 3 -s INCEPTION /workspace/images/mug.jpg
 
 # Inference should return the following
 Image '/workspace/images/mug.jpg':

TRITON_VERSION

@@ -1 +1 @@
-2.63.0dev
+2.62.0

build.py

@@ -71,14 +71,14 @@
 #
 
 DEFAULT_TRITON_VERSION_MAP = {
-    "release_version": "2.63.0dev",
-    "triton_container_version": "25.11dev",
-    "upstream_container_version": "25.09",
+    "release_version": "2.62.0",
+    "triton_container_version": "25.10",
+    "upstream_container_version": "25.10",
     "ort_version": "1.23.1",
     "ort_openvino_version": "2025.3.0",
     "standalone_openvino_version": "2025.3.0",
     "dcgm_version": "4.4.0-1",
-    "vllm_version": "0.10.1.1",
+    "vllm_version": "0.10.2",
     "rhel_py_version": "3.12.3",
 }
 
@@ -660,6 +660,10 @@ def pytorch_cmake_args(images):
         cargs.append(
             cmake_backend_enable("pytorch", "TRITON_ENABLE_NVTX", FLAGS.enable_nvtx)
         )
+        if target_platform() == "igpu":
+            cargs.append(
+                cmake_backend_enable("pytorch", "TRITON_PYTORCH_NVSHMEM", False)
+            )
     return cargs
 
 
@@ -1278,11 +1282,6 @@ def create_dockerfile_linux(
     pip3 install --no-cache-dir  grpcio-tools==1.64.0 && \\
     pip3 uninstall -y setuptools
 ENV LD_LIBRARY_PATH=/usr/local/tensorrt/lib/:/opt/tritonserver/backends/tensorrtllm:$LD_LIBRARY_PATH
-
-# There are some ucc issues when spawning mpi processes with ompi v4.1.7a1.
-# Downgrade to ompi v4.1.5rc2 to avoid the issue.
-RUN rm -fr /opt/hpcx/ompi
-COPY --from=nvcr.io/nvidia/tritonserver:24.02-py3-min /opt/hpcx/ompi /opt/hpcx/ompi
 """
     with open(os.path.join(ddir, dockerfile_name), "w") as dfile:
         dfile.write(df)
@@ -1461,14 +1460,6 @@ def dockerfile_prepare_container_linux(argmap, backends, enable_gpu, target_mach
             virtualenv \\
       && rm -rf /var/lib/apt/lists/*
 """
-    if "tensorrtllm" in backends:
-        df += """
-# Updating the openssh-client to fix for the CVE-2024-6387. This can be removed when trtllm uses a later CUDA container(12.5 or later)
-RUN apt-get update \\
-    && apt-get install -y --no-install-recommends \\
-        openssh-client \\
-    && rm -rf /var/lib/apt/lists/*
-    """
 
     if "vllm" in backends:
         df += f"""
@@ -1514,6 +1505,25 @@ def dockerfile_prepare_container_linux(argmap, backends, enable_gpu, target_mach
 ENV PYTHONPATH=/opt/tritonserver/backends/dali/wheel/dali:$PYTHONPATH
 """
 
+    if (
+        target_platform() not in ["igpu", "windows", "rhel"]
+        and "tensorrtllm" not in backends
+    ):
+        repo_arch = "sbsa" if target_machine == "aarch64" else "x86_64"
+        df += f"""
+RUN curl -o /tmp/cuda-keyring.deb \\
+        https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/{repo_arch}/cuda-keyring_1.1-1_all.deb \\
+      && apt install /tmp/cuda-keyring.deb \\
+      && rm /tmp/cuda-keyring.deb \\
+      && apt update -qq \\
+      && apt install --yes --no-install-recommends libnvshmem3-cuda-13 \\
+      && rm -rf /var/lib/apt/lists/* \\
+      && dpkg -L libnvshmem3-cuda-13 | grep libnvshmem_host.so | sed -e 's/libnvshmem_host.*//g' | sort -u > /etc/ld.so.conf.d/libnvshmem3-cuda-13.conf \\
+      && ldconfig
+""".format(
+            repo_arch=repo_arch
+        )
+
     df += """
 WORKDIR /opt/tritonserver
 RUN rm -fr /opt/tritonserver/*

deploy/aws/values.yaml

@@ -27,7 +27,7 @@
 replicaCount: 1
 
 image:
-  imageName: nvcr.io/nvidia/tritonserver:25.09-py3
+  imageName: nvcr.io/nvidia/tritonserver:25.10-py3
   pullPolicy: IfNotPresent
   modelRepositoryPath: s3://triton-inference-server-repository/model_repository
   numGpus: 1

deploy/fleetcommand/Chart.yaml

@@ -26,7 +26,7 @@
 
 apiVersion: v1
 # appVersion is the Triton version; update when changing release
-appVersion: 2.61.0"
+appVersion: 2.62.0"
 description: Triton Inference Server (Fleet Command)
 name: triton-inference-server
 # version is the Chart version; update when changing anything in the chart

deploy/fleetcommand/values.yaml

@@ -27,7 +27,7 @@
 replicaCount: 1
 
 image:
-  imageName: nvcr.io/nvidia/tritonserver:25.09-py3
+  imageName: nvcr.io/nvidia/tritonserver:25.10-py3
   pullPolicy: IfNotPresent
   numGpus: 1
   serverCommand: tritonserver
@@ -47,13 +47,13 @@ image:
     #
     # To set model control mode, uncomment and configure below
     # TODO: Fix the following url, it is invalid
-    # See https://github.com/triton-inference-server/server/blob/r25.09/docs/user_guide/model_management.md
+    # See https://github.com/triton-inference-server/server/blob/r25.10/docs/user_guide/model_management.md
     #  for more details
     #- --model-control-mode=explicit|poll|none
     #
     # Additional server args
     #
-    # see https://github.com/triton-inference-server/server/blob/r25.09/README.md
+    # see https://github.com/triton-inference-server/server/blob/r25.10/README.md
     #  for more details
 
 service:

deploy/gcp/values.yaml

@@ -27,7 +27,7 @@
 replicaCount: 1
 
 image:
-  imageName: nvcr.io/nvidia/tritonserver:25.09-py3
+  imageName: nvcr.io/nvidia/tritonserver:25.10-py3
   pullPolicy: IfNotPresent
   modelRepositoryPath: gs://triton-inference-server-repository/model_repository
   numGpus: 1

deploy/gke-marketplace-app/benchmark/perf-analyzer-script/triton_client.yaml

@@ -33,7 +33,7 @@ metadata:
   namespace: default
 spec:
   containers:
-  - image: nvcr.io/nvidia/tritonserver:25.09-py3-sdk
+  - image: nvcr.io/nvidia/tritonserver:25.10-py3-sdk
     imagePullPolicy: Always
     name: nv-triton-client
     securityContext:

deploy/gke-marketplace-app/server-deployer/build_and_push.sh

@@ -27,9 +27,9 @@
 
 export REGISTRY=gcr.io/$(gcloud config get-value project | tr ':' '/')
 export APP_NAME=tritonserver
-export MAJOR_VERSION=2.61
-export MINOR_VERSION=2.61.0
-export NGC_VERSION=25.09-py3
+export MAJOR_VERSION=2.62
+export MINOR_VERSION=2.62.0
+export NGC_VERSION=25.10-py3
 
 docker pull nvcr.io/nvidia/$APP_NAME:$NGC_VERSION